Valid_pct=1.0 error if trying to create cnn_learner for testset

Part 1 (2019)
1

Hey
i am trying to create a cnn_learner for my test set via from_csv. My plan is to set valid_pct=1.0 so that I can easily plot my toplosses using ClassificationInterpretation.from_learner(learn_train).

In the docs there is nothing about a limit for setting the % for the valid_set. But i am getting an error

Exception: Can't infer the type of your targets. 
It's either because your data source is empty or because your labelling function raised an error.

Can someone explain to me what this is about?

from Github

@classmethod
    def from_csv(cls, path:PathOrStr, folder:PathOrStr=None, label_delim:str=None, csv_labels:PathOrStr='labels.csv',
                 valid_pct:float=0.2, seed:int=None, fn_col:int=0, label_col:int=1, suffix:str='', delimiter:str=None,
                 header:Optional[Union[int,str]]='infer', **kwargs:Any)->'ImageDataBunch':
        "Create from a csv file in `path/csv_labels`."
        path = Path(path)
        df = pd.read_csv(path/csv_labels, header=header, delimiter=delimiter)
        return cls.from_df(path, df, folder=folder, label_delim=label_delim, valid_pct=valid_pct, seed=seed,
                fn_col=fn_col, label_col=label_col, suffix=suffix, **kwargs)
2

I assume you are trying to create a (Image)DataBunch Object with the from_csv factory method. Setting valid_pct = 1.0 means that you have no samples left to train your model. Often you set your validation set around 0.15 - 0.2 (which is the default), meaning taking 15 - 20% of your data as validation data. Try to set your valid_pct to 0.99 and a warning (not exception) should appear that your training set consists only of a few samples. My suggestion is to not set the valid_pct by your own so that your DataBunch uses the default value of 0.2. Then you are still able to utilize top_losses from your ClassificationInterpretation object.